Why Doesn’t the (Watch) Dog Bark?

1

Ph.D. in Applied Sociology and Methodology of Social Sciences

University of Milan-Bicocca

Department of Sociology and Social Research

Director of the Doctoral Programme: Professor Carmen Leccardi

Why Doesn’t the (Watch) Dog Bark?

Logics of Risk Regulation and

Management in the Italian Railway Sector

Ph.D. Candidate:

Supervisor:

3

A

I wish to express my deepest gratitude to my Supervisor Professor Maurizio Catino, the Directorof the Ph.D. Programme in Applied Sociology and Methodology of Social Science at the University of Milan-Bicocca, Professor Carmen Leccardi, all the Members of the Scientific Board of the Ph.D. Programme, Professor Marco Boniardi, Fabrizio D’Errico, and Franca Selvatici for the guidance they offered me during my research pathway.

I am very grateful to the Agenzia Nazionale per la Sicurezza delle Ferrovie – Italian National Safety Authority (NSA); the Direzione Generale per le

Investigazioni Ferroviarie – Italian National Investigation Body (NIB), and the

European Railway Agency and, particularly, to all the staff for their fundamental contribution to this research, without whom it would never have been possible.

During my Ph.D. I spent a research period at the Centre for Analysis of Risk and Regulation (London School of Economics and Political Science). I would like to thank Professor Michael Power and Professor Martin Lodge for giving me that opportunity.

I thank the participants of the Essex Summer School in Case Study Methods (August 2012), the brown bag seminar held by theCentre for Analysis of Risk and Regulation (July 2013), and the European Sociological Association Ph.D. Summer School ‘A Sociological Imagination for the 21st

Century’ (July 2014), for the helpful comments on the preliminary results of the analysis.

I am really grateful to Lisa Ariemma and William Besana for editing the thesis, and to Maksim Minkov for his graphics contribution.

A special thanks to Boscu, for the essential help he gave me in understanding railway language and technology.

During this research pathway, I had the opportunity of making really good friends who supported me and helped me in many different ways, among them, my Ph.D. colleagues Loris Mazzagatti, Sofia Pagliarin, and Emanuela Struffolino, and my Londoner friends, Layla Mettioui, Jessica Pereira, and Martha Poon.

Despite my travels in many different places, I know that when I come back home there is a strong local community in which, no matter what, I am always supported: the Susa Valley. Among my various hometown friends, I especially thank Federico Ambrosia, Danilo Favro (Manolo), Roberta Paviolo, Fabrizia Pinelli, Elena Rotondo, Edy Vavassori, and Samantha Vighetti.

5

C

INTRODUCTION ... 5

PART ONE. THEORETICAL FRAMEWORK: FROM RISKS TO ACCIDENTS THROUGH THE INSTITUTIONAL LOGICS PERSPECTIVE ... 25

1. RISK ... 28

1.1 The concept of risk ... 28

1.2 Risk management ... 55

1.3 Risk regulation ... 87

2. INSTITUTIONAL LOGICS PERSPECTIVE ... 103

2.1 The core meta-theory ... 105

2.2 Logics and the focus of attention ... 117

2.3 Institutional logics between theory and empirical research ... 121

3. ORGANISATIONAL ACCIDENTS ... 136

3.1 Main theories ... 137

3.2 Organisational accidents’ genesis ... 146

PART TWO. RESEARCH DESIGN: THE CASE OF THE ITALIAN RAILWAY SECTOR ... 153

1. A CASE STUDY RESEARCH DESIGN ... 157

1.1 Why a case study? ... 157

1.2 Case selection ... 162

2. DATA USED IN THIS DISSERTATION ... 166

2.1 Data collection ... 166

2.2 Data analysis ... 171

3. A MENTAL EXPERIMENT ... 177

PART THREE. DATA ANALYSIS: LOGICS OF RISK MANAGEMENT AND REGULATION IN THE ITALIAN RAILWAY NETWORK ... 183

1. LOGICS OF RISK MANAGEMENT AND REGULATION ... 187

1.1 Politico-economic level: risk-based logic ... 187

1.2 Inter-organisational level: cost-benefit, standard and possibility logics ... 201

2. LOGICS INTERPLAY AND DEGREE OF LEGITIMACY ... 335

PART FOUR. FINDINGS: COULD THIS NETWORK INTERCEPT AN ACCIDENT BEFORE IT HAPPENS? ... 343

1. FROM INTER-INSTITUTIONAL SYSTEM TO ORGANISATIONAL ACCIDENTS ... 344

2. LOOKING AT THE EVENTS WEARING THE INSTITUTIONAL LOGICS’ GLASSES ... 352

CONCLUSION ... 373

LIST OF FIGURES AND TABLES ... 385

ANNEXES ... 390

7

[…] I saw by the Inspector’s face that his attention had been keenly aroused.

“You consider that to be important?” he asked. “Exceedingly so.”

“Is there any point to which you would wish to draw my attention?” “To the curious incident of the dog in the night time.”

“The dog did nothing in the night time.”

“That was the curious incident,” remarked Sherlock Holmes.

[…] Before deciding that question I had grasped the significance of the silence of the dog, for one true inference invariably suggests others. The Simpson incident had shown me that a dog was kept in the stables, and yet, though someone had been in and had fetched out a horse, he had not barked enough to arouse the two lads in the loft. Obviously the midnight visitor was someone whom the dog knew well.

9

I

The object

The starting point of this dissertation is a specific type of event, the organisational accident. Organisational accidents are man-made (Turner, 1978; Turner and Pidgeon, 1997); not in the sense that they are intentionally driven, but are unintended consequences of human activities (Baldissera, 1998) such as, for example, the production of nuclear energy, the transport of goods and people, or oil drilling. These events are by definition extremely low in frequency and extremely high in magnitude. They differ from individual accidents because they happen to organisations, rather than to single individuals, and because of the high magnitude of their consequences (Reason, 1997). Thus, organisational accidents are extremely rare, but have important adverse and harmful consequences on health and the environment.1 Recent events such as the Costa Concordia accident, which occurred near Giglio Island on 13 January 2012; the Fukushima nuclear accident on 3 November 2011; or the Deep Water Horizon oil spill, which happened on 20 April 2010 in the Gulf of Mexico, are clear examples. Moving backward historically, the Chernobyl nuclear accident – 26 April 1986 – and the Bhopal gas leak accident – 2 December 1984 – are the most tragically famous examples with the greatest long-term consequences.2

During the last 30 years, the analysis of organisational accidents’ genesis has become a well-established and recognised field of study. This field has provided important insights into the genesis of organisational accidents identifying mechanisms/factors leading to the accidents within and from the point of view of the organisations in which the accidents happened (Turner, 1978; Reason, 1990; 1997; 2008; Vaughan, 1996; Turner and Pidgeon, 1997; Perrow, 1999; Snook, 2000; Catino, 2006; Downer, 2011). Thus, their focus has been at the organisational level. Let us take the example of the Deep Water Horizon oil spill mentioned above. The available studies analysing the accident genesis, focus on the mechanisms within British Petroleum – the company responsible for the management of the oil plant –

1

The term environment is used in a broader sense including all kinds of material damage.

2

10

which contributed to the accident’s occurrence. Examples of these mechanisms are: the presence of safety indicators referring exclusively to occupational safety, therefore individual rather than organisational accidents; a tendency to reduce costs without taking into account the possible consequences on safety; the existence of an organisational structure in which safety experts were responsible for setting standards but not for monitoring and enforcing them; and the reluctance of the management of the company to consider ‘bad news’ (Hopkins, 2012). However, organisations such as British Petroleum operate in a complex and vast environment. On the one hand, the possible dangerous and adverse consequences of human activities are the target of different organisations promoting and enforcing different ways of managing and regulating human activities in order to avoid, cope with, and/or handle the possible negative and unwanted outcomes of such activities. These organisations can be considered networks of private companies and public organisations interacting in order to avoid possible adverse and unwanted outcomes – thus at an inter-organisational level. On the other hand, the roles, responsibilities and regulatory approaches of the public and private organisations involved are shaped by the definition of legislative frameworks established at different levels of government such as the national, as well as the supra-national one – thus at a politico-economic level. Nevertheless, the inter-organisational and politico-economic levels rarely become the object of specific inquiries. For example, the available studies on the Deep Water Horizon oil spill do not analyse the network of public organisations in charge of monitoring and/or regulating the possible adverse outcomes of oil drilling, as well as their interactions with companies dealing with oil drilling – the inter-organisational level. Hence, organisations such as the US Environmental Protection Agency (EPA), the Mineral Management Service (MMS) – to which the US Department of Interior delegated its regulatory authority – and the US Coast Guard (USCG) are not objects of investigation. In the same way, available studies do not analyse the broader political and economic approach to regulating the negative outcomes of oil drilling – the politico-economic level. Therefore, the principles and ideas driving the definition of such a network of organisations, and specifying the roles and responsibilities in oil drilling monitoring and regulation, expressed in legislation such as the Outer Continental Shelf Lands Act (OCSLA), are not objects of inquiry.

11

organising in order to avoid, cope with, and/or handle possible adverse and unwanted outcomes of human activities, but they have different targets. Risk management examines the organisations dealing with a specific activity such as oil drilling, and fosters organisational processes and structures in order to deal with the possible negative outcomes of the on-going activity – the organisational level. In contrast, risk regulation targets the broader legislative frameworks organising human activities and fostering roles, and the responsibilities and regulatory approaches of the different organisations involved, in order to avoid, cope with, and/or handle possible negative outcomes – the politico-economic level. In the case of oil drilling in the United States, risk management would look inside companies such as British Petroleum. Risk regulation would look at the legislative framework such as, for example, the Outer Continental Shelf Lands Act (OCSLA), defining the network of organisations in which companies such as British Petroleum are embedded – the US Environmental Protection Agency (EPA), the Mineral Management Service (MMS), and the Coast Guard (USCG).

Available studies on risk management share a focus on the organisational level with organisational accident studies. In addition, such studies have a theoretical-normative nature. Thus, they highlight the ways in which organisations should organise themselves in order to avoid, cope with, and/or handle negative outcomes, rather than what the organisations actually do in order to avoid, cope with, or handle such outcomes.

12

and a description of the organisations involved in its implementation, such as the US Environmental Protection Agency (EPA), the Mineral Management Service (MMS), and the Coast Guard (USCG), and their formal roles and responsibilities (Dagg et al., 2011). However, they do not take into account the processes implemented, promoted and enforced by such public organisations, as well as the interactions between these organisations and the private ones such as British Petroleum; both involved in the regulatory process aiming to avoid, cope with, and/or handle the dangerous outcomes of oil drilling. Thus, risk regulation studies fail to address the way in which these public and private organisations offer a practical translation of such high-level regulation in their everyday activities: they do not analyse the politico-economic level ‘in action’. Furthermore, risk regulation studies do not consider the connections, possible contradictions and/or overlapping between different levels of government such as the national and the supra-national ones. An example in the US oil drilling case is the possible overlap, and/or contradictions between national and international agreements, such as the United Nations Convention on the Law of the Sea (UNCLOS) and the Outer Continental Shelf Lands Act (OCSLA).

More in general, the main gap that the risk management and risk regulation studies share is the absence of a clear link between risk management and regulation approaches/strategies and the available insights on the genesis of organisational accidents. Thus, the need ‘to understand the limitations of risk management [and regulation] approaches, that is to analyse situations in which they are helpful and when they might be counterproductive’ (Hutter, 2006: 220), is still not fulfilled.

The purpose of this study is to fill these gaps creating a bridge between these three fields of study – organisational accidents’ genesis, risk management and risk regulation – as well as closing in on the inter-organisational level whose analysis is still lacking. The aim is to focus on higher levels – inter-organisational and politico-economic – in contrast with organisational accidents and risk management studies, while maintaining a link with what is actually going on by observing such high levels ‘in action’. This contrasts with risk management – which is normative in nature – and risk regulation studies – which remain at a high level of abstraction. Our focus is on the inter-organisational level, thus on organisations in charge of regulating, monitoring and enforcing specific regulatory frameworks, but without losing the link with the possible negative outcomes of the regulated areas of human activities, as well as with society as a whole. Accordingly, the study aims to keep together two objects of analysis:

13

outcomes. This with a specific focus on the point of view of the organisations in charge of monitoring, and/or regulating the possible adverse outcomes of such areas of human activity at different levels of government – national and supra-national – but that are not involved in the core activity of the regulated domain. Thus, in the example of oil drilling in the US context, the network of organisations involved in management and regulation includes public regulators – the US Environmental Protection Agency (EPA), the Mineral Management Service (MMS), and the Coast Guard (USCG) – and private companies such as British Petroleum. In this network, we would consider the point of view of organisations such as the US Environmental Protection Agency (EPA), the Mineral Management Service (MMS), and the Coast Guard (USCG). Thus, organisations in charge of regulating and monitoring negative outcomes of oil drilling, but not dealing directly with oil-drilling activities and differing in this sense from companies such as British Petroleum;

 The organisational accidents, thus unintended events (Baldissera, 1998) happening to organisations rather than to single individuals, and characterised by extremely low frequency and extremely high consequences for health and environment (Reason, 1997). We can consider such events as a specific type of negative outcome that the network of organisations mentioned above should avoid, cope with, and/or handle.

By taking the point of view of regulators, we mean looking at:

 The politico-economic level ‘in action’: the way in which regulators interpret and translate the legislative framework, of which they are part, into practise;

 The inter-organisational level: the relationships within the network that go from the regulators to the regulated organisations, to other regulators, and to the phenomena they are in charge of regulating. Thus, the processes dealing with possible negative outcomes that the regulators promote, enforce, and monitor among the regulated organisations. The ways in which regulators coordinate themselves, as well as the ways in which they promote, enforce, and monitor processes among other regulators located at different levels of government. The processes such as information gathering and information analysis that regulators perform in order to gain awareness, and face the possible negative outcomes linked to the regulated area of activities, are also considered.

14

high degree of abstraction at the politico-economic level of study. Nevertheless, some of these processes can be considered as overlapping between the concepts of risk management and risk regulation. For example, if we look at the relationship between regulators and regulated organisations, the regulating organisations could promote and enforce specific processes of risk management fostering the way in which the regulated organise themselves in order to avoid, cope with, and/or handle the possible negative outcomes of their own activities. Such promotion and enforcement is part of their risk regulation strategies, but de facto expresses a specific approach to risk management. In addition, if we look at the relationship between regulators and the outcomes of the regulated activities, regulating organisations could develop their own processes of risk management directly within their own organisations, targeting the possible negative outcomes of the regulated area of human activity. Again, such processes are part of the risk regulation strategies of the regulating organisations, but identify a specific practical implementation of a risk management strategy. Returning to the example of oil drilling in the US, part of the regulatory activities performed by the US Environmental Protection Agency could be the promotion and enforcement of specific ways in which to manage risks in order to affect the way in which the possible adverse outcomes are managed within British Petroleum. In addition, organisations such as the US Environmental Protection Agency could look straight to the outcomes of British Petroleum’s activities in order to monitor the effectiveness of the risk management performed by such an organisation. Consequently, the US Environmental Protection Agency could perform risk management by itself, directly targeting the outcomes of the oil drilling activity. Thus, processes that can fit within both the concepts of risk management and risk regulation are taken into account.

The puzzle and the question

15

increasing focus by our (Western) societies on the avoidance of possible negative outcomes as a by-product of different areas of human activity (Hood and Jones, 1996; Coles et al., 2000; Hutter and Power, 2005; Taylor-Gooby and Zinn, 2006; Hutter, 2006; 2010; Rebora, 2007; Gephard et al., 2009; Minelli et al., 2009). Risk management and regulation have turned out to be a crucial public and political issue (Hood and Jones, 1996; Aven and Kristensen, 2005; Hutter, 2006; 2010; Taylor-Gooby and Zinn, 2006; Gephard et al., 2009). In addition, despite the variability affecting the definition of what harmful outcomes are, as well as the ways in which human activities should be/are managed and regulated in order to avoid, cope with, and/or handle such possible negative outcomes (Hood and Jones, 1996; Hood et al., 2001; Jasanoff, 2005a; 2005b; Rothstein et al., 2012), the number of public agencies dedicated to risk management and regulation has grown both at the national, as well as the supra-national levels (Braithwaite, 1982; 2003; Ayres and Braithwaite, 1992; Thatcher and Sweet Stone, 2002; Baldwin et al., 2012). Looking at the supra-national level, the creation of worldwide organisations in charge of coordinating the management and regulation of the possible negative outcomes linked to specific areas of human activities, such as the World Health Organisation, or the International Atomic Energy Agency, can be taken as examples (Scheytt et al., 2006). Consequently, a growing number of human activities have fostered complex networks of organisations – regulators and regulated organisations interacting at different levels of government – dedicated to risk management and regulation (Ibid.). In essence, a massive increase in attention, resources, and organisations dedicated to the risk management and regulation of human activities in order to avoid, cope with, and/or handle possible negative outcomes – organisational accidents among them – has emerged. Nevertheless, despite such an increase in attention, resources and the number of organisations, organisational accidents still happen. A look at recent news reports gives a considerable number of examples. The latest ones include: the derailment and explosion of an oil-transport train in Alabama on 8 November 2013; the Santiago de Compostela high-speed train derailment, which occurred 24 July 2013; the derailment and explosion of an oil-transport train in the town of Lac-Mégantic, Quebec on 6 July 2013.

16

The question does not directly consider organisational accidents’ genesis: why do accidents happen? In contrast with organisational accident studies, we do not aim to explain the aetiology of the accident. More specifically, organisational accident studies identify the mechanisms or factors within the organisations in which the accident happened, that acted as contributing factors creating a context prone to the accident’s occurrence (Reason, 1997). On the contrary, our aim is to explain why the regulators – organisations in charge of monitoring and/or regulating the possible negative outcomes of areas of human activity at different levels of government, but not involved in the core activity of the regulated domain – do not recognise the presence and the potential gravity of such mechanisms/factors. Thus, we are not linking or fostering the role of regulators as a contributing factor in the accident genesis. Instead, we question the possibility of such regulators becoming aware and acting in order to reduce or eliminate those mechanisms/factors, which can contribute to the accident genesis within the regulated organisations.

17

company itself started to encourage rather than discourage this dangerous practise. Basically, the inchino ritual was normalised (Vaughan, 1996): the constant violations of common sense safety boundaries were accepted and, given the absence of related accidents, such violations became a routine. Consequently, the dangerousness of the exercise, as well as the possible harmful consequences of such a practise, were underestimated or not perceived. Such a mechanism known by organisational accident scholars as the normalisation of deviance (Ibid.) can be considered an explanation of the organisational accidents’ genesis answering the question: why the accident happened?

From the point of view of the question: ‘Why doesn’t the (watch) dog bark?’, we would consider that the possible dangerous outcomes of the cruise ship’s activities are the target of a network of organisations. This network includes companies such as Costa, as well as different regulators such as the Corpo delle Capitanerie di Porto, dealing with navigation safety at the national level of government in Italy; and the European Maritime Safety Agency, at the European one. Consequently, our focus with regard to the inchino practise would be that such a custom was performed for almost thirty years before the Costa Concordia accident, without any awareness or intervention on the part of public organisations such as the Corpo delle Capitanerie

di Porto and the European Maritime Safety Agency. Hence, they are the very (watch)

dogs of maritime transport activities in charge of monitoring and regulating such activities in order to avoid possible negative and unwanted outcomes for health or environment. Our question aims to explain why, for example, none of those organisations recognised the dangerousness of the inchino practise; whether they were aware of such a practise; if they recognised the possible harmful outcomes linked to the practise; if they established clear safety limits – e.g. maximum mile limits – to shore approaches; monitored such practises; or monitored the consideration of such a ritual among the company’s risk evaluation processes.

Research pathway and main results

In order to answer the question – ‘Why doesn’t the (watch) dog bark?’ – we propose a two-step research design.

18

National Safety Authority (NSA); and the Direzione Generale per le Investigazioni

Ferroviarie – National Investigation Body (NIB), at the Italian level of government.

These three organisations can be considered the (watch) dogs of the Italian railway sector: organisations in charge of managing and regulating rail transport activities in order to avoid, cope with, and/or handle the possible negative outcomes of such activities, but not involved in rail transport operations. In order to understand the way in which the selected regulatory network works, we need a theoretical-analytical framework allowing us:

 On the one hand, to hold together different levels of government – national and supra-national – leaving room for contradictions and/or overlapping;

 On the other hand, to ensure an in-depth understanding of the processes, interactions and coordination strategies shaped by such organisations, as well as of the cultural and cognitive basis of such processes, interactions and coordination strategies.

The institutional logics theoretical-analytical framework (Alford and Friedland, 1985; Jackall, 1988; Friedland and Alford, 1991; Thornton and Ocasio, 2008; Thornton, Ocasio and Lounsbury, 2012) satisfying these requirements, frames the empirical analysis developed here. Institutional logics are ‘conceptual lenses’ (Allison and Zelikow, 1999) through which the regulating organisations see, interpret, and represent reality. The logic concept binds a set of symbolic components – categories and associated meanings, rationales, legitimate ends – and material ones – processes such as legitimated means to reach legitimated ends and structures – shaping and shaped by the organisation’s everyday on-going activities (Friedland and Alford, 1991). Thus, their identification and analysis allows us to understand the way in which the regulatory network functions, as well as the way in which such a network frames the negative outcomes and interacts in order to avoid, cope with, and/or handle them. The logics’ identification and description, and an understanding of their interaction are the product of the analysis of: different types of documents produced by the three regulating organisations – around 6,000 pages; 40 interviews conducted mainly with members of the three organisations; and observation of the everyday activities within each organisation – for a total of five months.

19

20

population of reference. The case selection follows some fixed attributes of the regulatory networks identified looking at previous risk regulation studies. More specifically, risk regulation studies have identified a general trend affecting the legislative frameworks promoted nowadays by (Western) societies in order to avoid, cope with, and/or handle the possible negative outcomes of human activities (Braithwaite, 1982; 2003; Ayres and Braithwaite, 1992; Osborne and Gaebler, 1992; Majone, 1994; 1997; 2002; Burton, 1997; Hood, 1991; Haufler, 2001; Hutter, 2001; James, 2001; Thatcher and Sweet Stone, 2002; Power, 2005; Estache and Wren-Lewis, 2010; Yeung, 2010; Aven, 2011; Baldwin et al., 2012). Such legislative frameworks shape some general attributes of the regulatory networks such as: the involvement of private organisations in regulatory activities (Haufler, 2001; Power, 2005; Aven, 2011); the presence of regulating organisations with a specific technical-scientific orientation, but out of the political arena and electoral pressures (Burton, 1997; Thatcher and Sweet Stone, 2002; Baldwin et al., 2012); and the existence of organisations located at different levels of government dealing with the same regulatory areas of activity (Estache and Wren-Lewis, 2010; Baldwin et al., 2012). We identify a case as close as possible to those attributes, thus some of the concepts and ideas developed here could in principle be useful to understand a broad and increasing population of cases.

21

Thus, following the ‘extreme case’ (Flyvbjerg 2006) strategy: an example of a regulatory network which performs particularly well has been selected. More specifically, the Italian case was chosen because:

 There is genuine commitment to the mission of ensuring a safe functioning of the regulated area of human activity, and various and effective risk regulation and management strategies have been developed pursuing such an end;

 There are a considerable amount of resources and skills dedicated to risk management and regulation within all the three organisations studied;

 There is a specific focus on the need to ensure regulator independence from the regulated organisations as well as from other regulators – the creation of three independent regulating organisations located at different levels of governments goes in this direction – and regulator accountability to society as a whole;

 There are structured processes of information gathered in place.

The extreme case strategy allows the effect of other intervening variables influencing the regulators’ activity to be minimised. Thus, minimising the interaction of other mechanisms that could affect regulators’ possibility to see warning signals, the logics’ focus of attention mechanism can be better identified.

The analysis points out that the regulatory network follows specific institutional logics. Looking at the politico-economic level in action – the way in which regulators interpret and translate the legislative framework of which they are part into practise – we identified a shared logic that we name the risk-based logic. With reference to the inter-organisational level – the relationships within the network that go from the regulators to the regulated organisations, to other regulators and to the phenomena they are in charge of regulating – we identified three different logics characterising the three regulating organisations’ approaches: the cost-benefit logic, prevalent within the European Railway Agency (ERA); the standard logic prevalent within the Italian National Safety Authority (NSA); and the possibility one prevalent within the Italian National Investigation Body (NIB). The three logics present different degrees of legitimacy, thus in their interplay one logic tends to prevail over others. Generally, the more legitimated one is the cost-benefit logic, thus the focus of attention shaped by this logic’s point of view tends to prevail during interactions and discussions.

22

on events that are relevant in order to intercept organisational accidents before they happen. More specifically, this study indicates how the same organisational processes, methods of reasoning, assumptions and principles shaping and shaped by regulators’ actions and decisions in order to manage the possible side-effects of the regulated area of human activity, tend to divert regulators’ attention from informational input that is potentially relevant in order to intercept an accident before it happens. Thus, it is not just a deviance or functional lacuna of the regulatory activity that can lead to an accident happening without any intervention by regulators, but it is the actual ‘normal’ functioning of the regulators’ activities that can prevent regulators from seeing events that are potentially relevant in intercepting an accident before it happens. Consequently, even in a perfect world of full resources, perfect and sound information and morally oriented behaviour, regulators’ activities are still affected by a bias of perspective that tends to select information focusing regulators’ attention on certain events instead of others. In conclusion, the study shows how despite the presence of regulating organisations, accidents do not happen because regulators are linked to regulated organisations exhibiting inappropriate relationships or conflicts of interest (Froud et al., 2004; Hirsch, 2003; Citron, 2003), adopt unethical or immoral behaviour (Mintzberg, 2004; Froud et al., 2004; Williams, 2004a, 2004b; Ghoshal, 2005), or lack sufficient or sound information (Vaughan, 1996; 2003; Reason, 1997). On the contrary, accidents can happen because standardised processes, accurate categories, precise definitions, scientific rigor and market rationality, can lead regulators to look away from events that are significant in organisational accidents’ genesis. In addition, even if other points of view are available, the high degree of legitimacy reinforcing such standardised processes, accurate categories, precise definitions, scientific rigor and market rationality, prevent other points of view from being considered. Thus, the (watch) dog does not bark because when the killer is approaching the victim, it is looking in another direction where it cannot see the killer nor hear him approaching.

Structure of the dissertation

23

and the contributions of previous studies that inform the case selection are described (Section 1.3). Chapter 2 presents the institutional logics perspective, as well as previous contributions dealing with the logics’ focus of attention mechanism. Chapter 3 is dedicated to organisational accident studies: organisational accidents’ main theories, and mechanisms/factors contributing to the accidents’ genesis identified by previous studies, are presented.

The second part deals with the research design. First, the case study research design is presented: justification of the choice to conduct a case study research design; explanation of the criteria used for the case selection (Chapter 1); and description of data used in the dissertation, as well as of the techniques used for data collection and analysis (Chapter 2). Then, the structure of the mental experiment, its limits and assumptions are presented (Chapter 3).

The third part is dedicated to data analysis. In order to specify the general background up on which this study is based, we propose a brief reconstruction of the history of the Italian railway sector identifying the main turning point in the legislative evolution of the management and regulation of possible negative and unwanted outcomes of railway transport activity. Then, we present the main evidence collected. First, the logics identified at the politico-economic level, as well as at the inter-organisational level are described (Chapter 1). Subsequently, the logics’ interplay and degree of legitimacy are discussed (Chapter 2).

The fourth part relates the identified logics to the macro and micro levels of analysis. First, the identified context-specific logics are located within society as a whole, and the contributing role of the various institutional orders such as the market, professions, and the state in shaping the context-specific logics, are examined (Chapter 1). An examination of the focus of attention effect – micro level effect – of the available logics follows. Then, by comparing the events up on which the logics focus regulators’ attention, and the events previous empirical research identifies as relevant in organisational accidents’ genesis, we can finally answer the question ‘Why doesn’t the (watch) dog bark?’ (Chapter 2).

25

P

.

T

:

F

26

This part aims to present the theoretical framework that has guided the development of the study. The structure of the chapters follows the three main fields of reference this dissertation looks at: the ‘risk province’, the institutional logics perspective and the organisational accident field.

With reference to the ‘risk province’ chapter, the aim of each section is to discuss the definition of risk (Chapter 1, Section 1.1), risk management (Chapter 1, Section 1.2), risk regulation (Chapter 1, Section 1.3); and to describe the main contributions of each field of reference. Before looking closely at each field of reference, a few words on the way in which these fields crop together are needed. With reference to the distinction between risk management and risk regulation, as already mentioned, the boundary between the two fields has become more and more blurred over the years. There are two main reasons for the overlapping: on the one hand, a change in regulatory strategies (Braithwaite, 1982; 2003; Ayres and Braithwaite, 1992; Osborne and Gaebler, 1992; Estache and Wren-Lewis, 2010; Yeung, 2010; Aven 2011; Baldwin et al., 2012); on the other, an effort made by scholars for the development of an interdisciplinary field of study (Hutter, 2006). The change in regulatory strategies indicates a move away from the control-command role of the state as well as a strong involvement on the part of the regulated organisations in regulatory activities. (See Part 1, Chapter 1, Section 1.3.) The distinction between the role of the state and the role of the regulated domain becomes less defined. Consequently, the distinction between regulation and management loses relevance as well. With reference to the scholars’ role, Hutter (2006; 2010) describes it as ‘a concerted effort to delineate a new interdisciplinary area of risk regulation studies which would bridge regulation and risk management studies.’ Such a concerted effort has slowly led to the definition of a more hybrid field of study still in development. Despite such blurring boundaries, in this chapter the two fields are dealt with separately. On the one hand, risk management (Section 1.2) and regulation (Section 1.3) remain two historically different fields of study. On the other, these fields’ statements have played a different role in the definition of the research design. More specifically, the risk regulation studies available played a crucial role in the case selection; in contrast, available work on risk does not played a role in the definition of the criteria for case selection. However, during the description it is possible to notice some overlapping.

27

theoretical analytical framework structuring the empirical analysis (Part 1, Chapter 2) goes in this direction. The institutional logics perspective effectively allows a bringing together of risk management and regulation, in a dialectic way, modelling these two elements in a multilevel analytical framework.

28 1. RISK

1.1 The concept of risk

This section aims to explore the concept of risk. First, by analysing some fields of human activity in which the concept is used, we identify the different meanings as well as the various phenomena bound under such a concept in society nowadays. Then, we present the main anthropological and sociological theories on risk, highlighting the way in which they differ from one another, as well as from the ‘technical-scientific’ approach to risk. Such an analysis allows us to understand the main strengths and weaknesses of the available theories and establish the scenario in which this study aims to locate itself. In conclusion, the definition of risk used as a starting point of this dissertation; the kind of phenomena this study aims to focus on using the term risk; as well as the position this study aims to maintain with respect to previous anthropological and sociological theories, are specified.

Risk: the meaning

As Hood and Jones (1996: 2) underline ‘there is not a clear and commonly agreed definition of what the term “risk” actually means.’ The Oxford English Dictionary3 defines risk as:

 A situation involving exposure to danger;

 The possibility that something unpleasant or unwelcome will happen;

 A possibility of harm or damage against which something is insured;

 The possibility of financial loss.

The origin of the term remains unknown, but it is first used around the middle 17th century. The etymological root is the French risqué – noun – and risquer – verb – and the Italian rischio – noun – and rischiare – verb (Ibid.). Despite the uncertain origin of the term, as well as its overlapping with other neighbouring terms such as hazard, uncertainty, or fate, the idea of risk was already conceived during the Roman Empire. For example, in 215 B.C. Livy (XXIII, 48 and 49) describes a government guarantee, requested and obtained by some goods producers and sellers for transporting goods from Rome to Spain, in order to supply the legions deployed there

3

29

at that time. More specifically, the agreement referred to the uncertainty of maritime transport: in case of loss of the transported goods due, for example, to adverse weather conditions, the value of the lost goods would be reimbursed by the Empire (Trenerry, 2009). Thus, the Empire took the risk – the possibility of losing the transported goods during the trip – covering the cost of the goods even if they were lost or damaged during transportation. Moving forward historically, in the Middle Ages, maritime trade into uncharted waters exhibits a similar conception of risk (Oppenheim, 1954; Ewald, 1991; Luhmann, 1993; Giddens, 1999).

A first formal definition of the term appears at the beginning of the 20th century. The risk panorama has been considerably shaped by such a mainstream definition, which, from its origin, (Knight, 1921) has deeply influenced further development of the identification of risks and their management. Following such a definition risk is the product of two variables:

 Probability of the event: the estimation of the number of times the event could happen given a specified period of time;

30

Fields Meaning Neighbouring concepts Bound phenomena Insurance Risk is an unavoidable danger

(irreducible – fate, destiny), but with quantifiable monetary losses if we take into account

groups of

individuals/organisations as a whole

Risk vs. Fate Risk = Probability of an event and monetary loss linked to the

event (estimated only when referring to classes of events

affecting groups of people/organisations, not estimable for single individuals

or organisations)

Negative events clearly linkable and affecting groups of individuals or

organisations (e.g., negative event: car crash

affected group of individuals: car drivers)

Business and finance

Risk as intrinsic component of any enterprise with a positive or negative meaning according

to the variance affecting the expected outcome

Risk vs. Uncertainty Risk = Probability (measurable uncertainty) per magnitude. Risk

bound phenomena that can be linked to phenomena of the same

kind that happened in the past

Monetary outcomes of the business or financial activity Technology, health care, safety at work

Risk has to do with the reliability of human activities: the possibility of negative and unwanted outcomes for health or environment rising from the side-effects of intentionally

driven human activities.

Risk vs. Hazard: Hazard becomes risk if it enters into a process aiming to estimate – qualitative or quantitative estimation of probability and magnitude (formal mainstream

definition not used in all the mentioned fields) – deal with and

control it Risk vs. Fate:

Outcomes are not completely out of human control and can in principle be avoided, coped with

and/or handled

Side-effects of human activities (e.g. nuclear power plant: accidental release of radiation into the environment; health care: accidental confusion

of one drug with another leading to the death of a

patient) Natural hazard and national security

Risk identifies the possible consequences of an external threat (nature or human activity

expressively aiming to create damage to society). Risk is

conceived as, in principle, manageable, not completely out

of human control National security: relational

nature of the phenomena

Risk as Hazard/Threat: The focus is on the consequences

of the events on health and environment.

Risk vs. Fate or Destiny: despite the external nature of the threat, human activity can interact with the dangerous event in order to avoid, cope with and/or handle it.

Consequences of external threats: consequences of natural phenomena such as landslides, earthquakes or tornados; consequences

of terrorist attacks.

31

Looking at the insurance field, in which the maritime trade’s guarantees and agreements mentioned above can be considered historical precursors, the term risk identifies the monetary losses as consequences of negative events that can affect individuals or organisations. As insurance agents are known to say: insurance cannot change destiny, but can limit the monetary damage associated with it (Pellino at al., 2006).4 On the one hand, given the unavoidable nature of the events, risk is linked with fate or destiny; but, on the other hand, the global amount of monetary losses carried out by fate or destiny can be estimated and quantified. Thus, the quantification of the monetary losses linked to the negative events is what characterises the definition of risk and distinguishes risk from fate or destiny. For example, we can estimate the number of car crashes happening every year in a specific location, as well as the monetary costs associated with such car crashes, but we cannot estimate exactly when and to whom they will happen. In addition, the insurance risk concept binds phenomena that can be associated with a specific class of individuals – e.g.: car insurance for human beings that drive a car, life insurance and health care insurance for human beings in different health conditions, such as smokers and non-smokers – or organisations – e.g. public liability insurance and environmental damage liability insurance for companies, dealing with different types of activities, such as nuclear energy production, or building carousels. Following the insurance approach, risk can be quantified, but only taking into account the class of individuals or organisations as a whole (Ewald, 1991). Thus, with a fixed class of individuals/organisations of reference, there is a component of certainty that relates to the total amount of adverse events and the costs linked to the consequences of such events in a given period of time, as well as a specific spatial location. A component of uncertainty related to establishing when, during a selected period, and to whom, within the selected class of individuals or organisations, the event will happen. The formalisation of a mathematics of probability – dated generally at 1713, the year of publication of Bernoulli’s Ars conjectandi (Hacking, 1975) – constituted an important turning point in the estimation of risk (Taylor-Gooby and Zinn, 2006). Thus, the estimation of potential monetary losses as consequences of events affecting specific groups of individuals/organisations has become more and more

4

32

sophisticated. Thereafter, the development of actuarial science since the end of the seventeenth century has led to the implementation of increasingly advanced statistical-mathematical models (Haberman and Sibbett, 1995; Lewin, 2003). Consequently, the quantification of the number of negative events and of the amount of monetary damage carried out by such events, as well as the characteristics defining the various targeted classes of individuals or organisations to which a specific amount of events and monetary damage is associated, has become more and more formalised.

33

uncertainty concept binds phenomena, characterised by non-measurable uncertainty: phenomena that are one-off ones, not linkable to previous phenomena of the same kind. Using Knight’s (1921: 233) words:

The practical difference between the two categories, risk and uncertainty, is that in the former the distribution of the outcome in a group of instances is known (either through calculation a priori or from statistics of past experience), while in the case of uncertainty this is not true, the reason being in general that it is impossible to form a group of instances, because the situation dealt with is in a high degree unique.

Since the beginning, Knight’s idea of probability has been criticised by scholars sustaining a subjectivist interpretation of probability (e.g. Ramsey, 1931; de Finetti, 1937; Savage, 1954), in which the strong boundary between risk and uncertainty identified by Knight loses relevance. Nowadays, the Bayesian theory of probability offers a formalisation of the subjectivist element affecting the probability estimation (e.g. Aven and Kristensen, 2005; Aven, 2011).5 Despite such developments, the deterministic and objectivistic assumptions on which Knight’s definition relies have been broadly accepted and have spread into other fields of study – for example, engineering technological risk management – currently representing the mainstream approach to risk definition and management.

While the business and financial fields include the variability of intended outcomes of business and financial enterprise under the concept of risk, in other fields such as technology-based activities, health care, or safety at work, the concept of risk identifies the unwanted outcomes of such activities as by-products of the intended ones. In this case, the term risk refers to the possibility of negative and unwanted outcomes for health or the environment arising from the side-effects (Merton, 1936; 1940; 1968) of intentionally driven human activities: consequences of actions that contrast with the intentional ends of such actions. Thus, risk has to do with the reliability of the performed activities rather than, as the business or financial definition of risk indicates, with the efficacy of such activities. For example, looking at technology-based activities, the end of a nuclear power plant is to produce energy, but such an activity runs the risk of leading to unwanted negative side-effects such as

5

34

the release of radiation into the environment: the possibility of a radiation release – side-effect of the intentional end of a nuclear plant which is the production of energy – is part of the technological risk to which a nuclear power plant is exposed. Looking at health care, the purpose of a nurse’s working activity in a hospital is to take care of ill people, for example through the delivery of pharmaceuticals, but as a possible side-effect of the process of care, the nurse could accidentally confuse one pharmaceutical with another one – e.g. potassium chloride instead of sodium chloride – and such an error could lead, in the worst case, to the death of a patient. The possibility of delivering an incorrect pharmaceutical that leads to the death of the patient is an example of the risk associated with health care activity, usually named clinical risk. Looking at work safety, the end of a press-worker is, for example, to shape a piece of steel through the activation of the press. As a side-effect of such activity, the worker runs the risk of accidentally activating the press while he is placing a piece of steel inside the press crushing his hand under the press. The possibility of crushing his hand is an example of the risk associated with the working activity usually named occupational risk. Other examples of risk as side-effects of human activity are: attacks on children from convicted paedophiles released from prison; injuries and death from motor vehicles on local roads; adverse health effects from exposure to pesticide residue both in food as well as in drinking water (Hood et al., 2001).6 The definition of risk as a side-effect differs from business, financial and insurance risk. On the one hand, even if the economic losses following the negative events are in principle quantifiable, the accent is usually on the negative consequences on health and the environment rather than on the monetary value of such consequences. On the other hand, risk is seen as something not completely out of human control, such as fate or destiny, but there is a possibility to interact with the on-going activities in order to avoid, cope with, and/or handle the possible negative events. Mainstream technological and health care risk management usually distinguishes between hazard and risk, in a way similar to Knight’s (1921) distinction between uncertainty and risk: hazard is the possible identified threat, the identified threat becomes a risk once a probability – defined quantitatively or qualitatively – and a magnitude are associated with it (e.g. British Health and Safety Authority).7

6

Hood et al. (2001), in analysing the variability of the risk regulation regime within the UK, place their attention on the phenomena mentioned above enlarging the targets of scholars’ analysis, which previously focused mostly on technological, occupational or clinical risks, onto other areas of human activity similarly affected by risk as a side-effect of such activities.

7

35

Despite the different bound phenomena, many concepts and management strategies developed within the business and finance fields have been borrowed from, and adapted, by technology-based activities and health care fields.

Natural phenomena as well as national security are other fields in which the term risk is employed. Such fields differ from the business, finance, and side-effect ones because of the external nature of the threat. The insurance risk definition is partially bound by these phenomena, however. Nevertheless, despite the external nature of the threat, in this case, unlike the insurance case, risk is conceived as manageable, in principle, in order to avoid, cope with, or handle it; thus, clearly separating it from the contiguous concepts of fate or destiny. Natural phenomena such as landslides, earthquakes or tornados are not in principle dangerous but, within the human environment in which they happen, they can have important negative consequences on health or the environment. Such a focus on the consequences of events, rather than on events in and of themselves, leads to the use of the term hazard instead of, or as a synonym of, risk (Hood and Jones, 1996). National security risk, such as a terrorist attack threat, is the intended outcome of other activities that national intelligence has to face. In this sense, a national security threat shares the external nature of danger with the negative consequences of natural phenomena. However, national security threats differ from natural phenomena for the relational nature of any intervention (e.g. Allison and Zelikow, 1999). For example, if a terrorist attack threat is identified and control measures are implemented in order to avoid it, the terrorists planning the attack can change their strategy as well in order to bypass the adopted control measures. Despite such differences, threat identification and assessment show many contact points with other fields of human activities, as well as with the identification of the possible negative consequences of natural phenomena. In addition, as a side-effect risk, the definition of risk binding possible negative consequences of natural phenomena, as well as of terrorist attacks, focuses on the negative consequences on health and the environment, rather than uniquely on the quantification of the monetary losses linked to the consequences of such events.

36

 The monetary loss of negative events – linked in some respect to fate or destiny – associated with specific classes of individuals or organisations – insurance risk;

 The positive or negative variance of the outcomes of any business or financial enterprise quantifiable through the product of probability – distinguishing risk from uncertainty – and the magnitude of the outcome – business or financial risk;

 The possibility of side-effects of human activities leading to negative consequences on health or the environment, thus the identification of probability and magnitude of such events distinguishes risk from hazard – technological risk, clinical risk and occupational risk;

 The possible consequences – a focus on the consequences leads to the use of both the terms risk and hazard – of external phenomena such as natural phenomena or a terrorist attack – environmental risk and terroristic risk. The variability affecting the different examined fields provides an initial idea of the complexity surrounding the risk concept. Since the end of the 1980s, social sciences have started looking closely at the concept of risk, underlining the socially constructed nature of such a concept, and contrasting it with the objectivistic assumptions upon which the mainstream ‘technical-scientific’ approaches rely. The next section summarises the main social science theories on risk, stressing the differences between the ontological and theoretical views on which such theories are based. Such ontological and theoretical plurality reflects the complexity surrounding the use of the concept in the different fields of human activities mentioned above.

Risk and social science: an analytical map

Risk is objective and scientifically knowable. Risk is subjective and socially constructed. Risk is a problem, a threat, a source of insecurity. Risk is a pleasure, a thrill, a source of profit and freedom. Risk is the means whereby we colonise and control the future. Risk society is our late modern world spinning out of control (Garland, 2003: 49).

37

influence the way in which scholars deal with the definition of risk, and build on theories. Furthermore, those assumptions ‘lead to different research programs as well as different interpretations of the results’ (Zinn, 2008: 2). This chapter aims, to map the main social scientists’ theoretical works on risk, locating them with reference to their ontological and theoretical starting points. In order to offer an analytical synthesis of the main theoretical approaches. We propose a Cartesian diagram (see Figure 1) based on two main dimensions: the theoretical dimension – x-axis – and the ontological dimension – y-axis. Both the axes are conceived as a continuum along which the various approaches are located. The theoretical dimension – y-axis – draws attention to the different accents placed by scholars on individuals or on society. More specifically, the top-down pole underlines a focus on social structures, institutions or functions and a deterministic role of society in shaping individuals’ perceptions and behaviour. In contrast, the bottom-up pole focuses on individuals’ perceptions and behaviour weakening the role of social structures, institutions or functions. The various approaches to the study of risk can be placed on a continuum, moving from a micro level, focusing on the individual, to a macro level, focusing on society. The ontological dimension – y-axis – refers to the nature of risk and can vary from realism to constructivism.8 Realism conceptualises risk as a real object, having an objective existence and being objectively knowable. Constructivism emphasises the role of culture and society in risk identification and definition. This conception is well represented by Ewald’s famous statement (1991: 199): ‘nothing is a risk in itself; there is no risk in reality. But on the other hand, anything can be a risk.’ (Emphasis in the text.) The various theories can be located along a continuum leading from risk as real to risk as social construction.

Figure 1 shows how the main approaches of social scientists on risk are situated on the Cartesian diagram taking into account both the theoretical as well as the ontological continuum. This analytical synthesis has the advantage of offering an overview of a diversified set of theoretical frameworks. However, ‘every classification is a trade-off between analytical rigor and empirical appropriateness’ (Catino, 2013: 220), and involves a loss of information. Thus, classifications have the advantage of allowing us to clearly locate this dissertation’s approach from the previous contributions – see next section – although they imply a simplification and

8

38

polarisation of the analysed theories. Let us look closely at the main sociological and anthropological contributions contextualising them through the diagram and in relation to the approaches of other disciplines. Anthropological and sociological contributions arise as a critique of ‘technical-scientific’ risk identification and analysis – engineering, medicine, epidemiology – and, looking at social sciences – psychometric, psychology, and economics – are all located in the fourth quadrant of the Cartesian diagram. Those disciplines look at risk as an objective and measurable phenomenon, and their main research interests are ‘identification of risks, mapping their causal factors, building predictive models of risks relations and people’s response to various types of risk and proposing ways of limiting the effects of risks’ (Lupton, 1999b: 2). Social sciences closest to the technical-scientific approach – e.g. psychometric, psychology and economics – are usually focused on individuals seen as acting independently from the social, cultural and historical context in which they live. In contrast, anthropological and sociological analyses of risk bring the social and cultural dimensions on the scene, and highlight the role of cultural and social elements in risk identification and definition. The main anthropological and sociological theories on risk can be grouped into five main categories: cultural symbolism, risk society, systems theory, governmentality and edgework (Lupton, 1999b; Arnoldi, 2009; Zinn, 2009). As a first approximation, the diagram shows a clear distinction between theories giving a dominant role to individuals – third quadrant – and theories focusing on structures, institutions, or functions – second quadrant. With regard to the ontological status of risk, even if all theories refute the definition of risk as an objective and measurable entity – any theory located in the high part of the first and fourth quadrant – we can see the more blurred positions of the different theories. A description of the five identified theoretical approaches focusing on the main scholars’ works within each category follows.

39

Figure1: Social sciences and risk – an analytical map (My elaboration.) Cultural symbolism

40

potential dangers that surround them and interact so as to concentrate only on selected aspects’ (Ibid.: 9).

With respect to the theoretical dimension – the x-axis – Douglas’s approach to risk is clearly situated near the top-down pole. More specifically, functionalism and structuralism are the frames of reference of her analysis: ‘for an anthropologist the human factor would mean the general structure of authority in the institution’ (Douglas, 1992: 12). Following such an approach, individual perceptions of risk and the future are shaped, or determined, by the characteristics – mode of social organisation and membership – of the community of belonging: groups, organisations, institutions or society as a whole. Cultural beliefs guide the way in which people frame risk and the future, so the perceptions and the definitions of risk are culturally shaped. Culture is just a product of social institutions and, at the same time, those institutions are supported by the culture they form. For example, the grid and group typology (Douglas, 1978; Douglas and Wildawsky, 1982, Thompson and Wildawsky, 1986) well represents Douglas’s approach. The group dimension identifies the degree of commitment of the members of a group – community, organisation, or society – and the degree of demarcation of the borders separating the members from broader society. Instead, the grid dimension focuses on the structuring degree of the group – hierarchy, and chain of command. From those two dimensions, Douglas identifies four main types of groups: hierarchy, entrepreneurs, sectarian, and excluded. Each type of group reflects different values outlining a specific conception and expectation about risk, the future, and the environment.

41

(Douglas and Wildavsky, 1982: 14). The moral and political implications in risk definition are also rejected in terms of the processes of objective scientific inquiry on risk:

In a context of scientific inquiry, an objective statement is arrived at by standardised techniques; the inquiry can be replicated and under the same conditions will reproduce the same answer. However objective the process, the interpretation is not guaranteed right by objectivity in the research design (Ibid.: 72).

Douglas’s historical analysis of risk indicates continuity between modern and pre-modern society. She argues against the distinction between pre-modern and rational ways of coping and managing risks, and pre-modern and irrational ones (Douglas, 1986). She underlines the continuity between traditional and modern responses to risk bringing risk back to structural factors: ‘The difference is not in the quality of knowledge but in the kind of community that we want to make […] or […] the community that technology makes possible for us’ (Douglas, 1992: 10). Douglas’s work is of great value for social sciences since she introduces the cultural dimension into risk analysis, but it is subject to the same criticism traditionally addressed to functionalism and structuralism. In brief, there is no room for individual agency, therefore it is not clear how social changes could be possible and, on a global scale, a static vision of cultural processes emerges.

Risk society

Risk society theory’s most famous contributions are the works of Urlich Beck (1986; 2007).9 Beck aims to develop a general theory of our society, focusing on the turning points – end of 1960s – between industrial society and what he has called risk society. Risk is seen as a distinctive element of society today. With regard to the theoretical component, risk society theories are located close to the top-down pole. These theories focus on macro-structural elements influencing radically modern, specific and elevated concern about risk. The main institutions of modernity – government, industry and science – are indicated as the main producers of contradictions and risks. The scientific statements regarding risk lost their

9_{Antony Giddens’s (1990, 1991) work on risk is fairly close to Beck’s. In this summary, I chose to}